PFs are often characterized by small fault throw (Figure 3A), small lateral extension and various displacement value (from a few meters to over 100 m), mainly developed in shallow-buried strata which are over 100 m thick and deposited after the Middle Cretaceous period (mainly Tertiary, Figure 3B, Figure 4B; Cartwright, 2007; He et al., 2010; Berndt et al., 2012; Ostanin et al., 2012; Li et al., 2020). They have neither dominant strike nor dominant tendency, existing as networks of minor normal faults (Watterson et al., 2000; Alrefaee et al., 2018; Figure 2). The throws of PFs developed in most of the PF areas are smaller than 50 m, such as in the Lower Congo Basin (10–20 m, Morgan et al., 2015), in the Hammerfest Basin (5–20 m, Morgan et al., 2015), in the Qiongdongnan Basin (10–40 m, Han et al., 2016), and in the Near Farafra Oasis, Egypt (2–3 m, Tewksbury et al., 2014). In addition, some PFs with larger throw of >100 m have also been found such as in the Faroe-Shetland Basin (Bureau et al., 2013). The extending length of PFs in most of the PF areas is generally smaller than 2 km, such as in the Northern Carnarvon Basin (70–620 m, Alrefaee et al., 2018), in the North Sea Basin (1,000–2000 m, Lonergan et al., 1998a), and in the Pearl River Mouth Basin (100–1400 m, Sun et al., 2012).

With the increase of burial depth, the dip angle of PFs decreases gradually, interpreted to be caused by compaction and flattening of the fault planes due to increasing vertical stress (Lonergan et al., 1998b; Stuevold et al., 2003; Shoulders et al., 2007). For example, some PFs from offshore Norway appear to have relatively steep dip angle of ∼80° at the shallower level, but they show strongly listric at deeper depth with comparatively smaller dip angle of ∼60° (Cartwright, 2011). PFs are layer-bound deformation structures mainly formed in shallow soft sediments of marine settings. Most of the PFs discovered worldwide are located at depth smaller than 1,500 m below the seabed (mbsb) with only a small proportion of PFs occur at depth ranging from 1,500 to 2000 mbsb, at which depth the early stages of compaction probably caused the initiation of PFs (Figure 4A). At even deeper depth, the PF-bearing layer experienced strong diagenesis which makes the PFs difficult to be imaged in seismic. What is more, small sized PFs have been found in a few meters below the present-day lakebed in the Western Lake Superior, with a thickness range of 10–15 m and fault throw of < 1 m (Cartwright et al., 2004). Two outcrops of PFs have been observed in the Jurassic layers in the Arches National Park in south-eastern Utah and the Cretaceous layers in the Near Farafra Oasis, Egypt (Tewksbury et al., 2014 Antonellini and Mollema, 2015). The layer-bound PFs are mainly developed in marine claystone, and therefore the thickness of the PF-bearing strata is mainly controlled by the sediment distribution deposited in certain sedimentary environment (Cartwright and Lonergan, 1997; Goulty, 2008). Most PFs around the world were developed in layers of < 700 m thick, with quite a few occurring in layers of only several meters thick (Table 2; Figure 3B). For example, the outcrops of PFs was observed in Arches National Park in south-eastern Utah, the layer thickness of which is < 2 m (Antonellini and Mollema, 2015).

PFs are usually found in layer-bound successions consisting mainly of fine-grained sediments (e.g., mudstone, shale, claystone, or carbonate chalk), which may be related with the genesis mechanism of PFs (Andresen and Huuse, 2011; Cartwright, 2011; Yang et al., 2017; Alrefaee et al., 2018; Ireland et al., 2021). “Syneresis” due to spontaneous contraction of colloidal sediments and “shear failure” caused by mineral-specific dissolution have been widely accepted to explain the PF formation mechanism in fine-grained sediments (Han et al., 2016; Li et al., 2020). Besides, small outcrops of PFs have been found in the Entrada sandstone strata of Middle Jurassic in the Arches National Park in south-eastern Utah (Antonellini and Mollema, 2015). Since sandstone are relatively coarser and permeable, genesis mechanisms such as “syneresis” and “shear failure” cannot explain the PF formation in sandstone layers. The “density inversion” was proposed to account for the PF formation in the Entrada Sandstone (Antonellini and Mollema, 2015). The over-pressured sequences (Carmel Formation) is overlain by normally pressured layer (Entrada Sandstone), which caused accompanied shale flowage and hydrofracturing therefore leading to the formation of PFs in sandstones. In addition, the geometry, distribution and intensity of PFs are mainly determined by lithological features such as grain size, hardness, diagenesis, clay mineralogy and organic matter content (Dewhurst et al., 1999 ab; Hansen et al., 2004; Turrini et al., 2017). When the PF-bearing layer is homogeneous in thickness and lithology, faults cross cut each other uniformly at an angle of 60° (Cartwright, 2011; Li et al., 2020). Otherwise, the geometry of the PFs is featured by irregular polygons when the PF-bearing layer is heterogeneous in thickness and lithology, which represents the situation of most global PF cases.

Although most PFs occur in fine-grained sediments, their formation and characteristics are also affected by nearby coarse sand bodies density (Victor and Moretti, 2006; Shoulders et al., 2007; Liu et al., 2008; He et al., 2010). In the south-east area of the Sanzhao sag, the PFs are well developed in areas of smaller sand/layer thickness ratio and poor physical property. On the contrary, the PFs are sparsely developed in the northern area of Sanzhao sag where is featured by larger sand/layer thickness ratio and good physical property (He et al., 2010). The sand bodies overlying the PFs favour the discharge of fluids from the PF-bearing mudstone layer, inhibiting the formation of overpressure and therefore causing sparsely developed PFs.

According to previous studies, the presence of PFs has been proposed to play a role in both favouring fluid migration and as a fluid source during the early stage of PF generation (Hustoft et al., 2007; Sun et al., 2010; Tewksbury et al., 2014). Many researchers commonly believed that PFs facilitate fluid migration by generating more failures and destroying the integrity of fine-grained layers, and regarded PFs as dominant channels for fluid migration and re-migration (Ding et al., 2013; Seebeck et al., 2015; Han et al., 2016). For example, Laurent et al. (2012) has proposed that the PFs in Gjallar Ridge offshore Norway can act as preferential fluid migration pathways when they are connected closely. Alrefaee et al. (2018) interpreted the PFs in Northern Carnarvon Basin in Australia as the fluid migration pathways based on seismic imaging of the discontinuity of the fine-grained sediments. In the Sanzhao Sag of the Songliao Basin, the oil source comparison of hydrocarbon reservoirs above and below PFs shows that they were from the same source rock, and this was regarded as the evidence that PFs acted as the fluid migration pathways (He et al., 2010; Ding et al., 2013). According to Sun et al. (2010), the development of PFs in the fine-grained Huangliu Formation and Meishan Formation in Qiongdongnan Basin destroyed the sealing property and increased the permeability, resulting in the upward secondary migration of oil and gas from the lower hydrocarbon reservoir.

Considering the dewatering process during the initial period of PF formation, the PF-bearing layer may be a fluid source where fluids were expelled to nearby sediments. For example, the sediment dewatering during PF generation in the Kai Formation and the underlying Brygge Formation in Mid-Norwegian Margin were regarded as responsible for providing the fluid source for the gas hydrate formation in the shallower Vøring margin (Berndt et al., 2003; Hustoft et al., 2007). PF generation is commonly associated with underlying troughs and overlying incised valleys, and it was proposed that the fluid produced by dewatering in the early stage of PFs formation may escape and migrate upward causing the troughs due to related consequent volume loss and subsequent collapse (Sun et al., 2010). The overlying incised valleys were caused by fluid escape which was expelled from underlying PF-bearing mudstone layer (Figure 5A).

Sediment compaction and pore water drainage occur during PF formation, resulting in the formation of pockmark in Canterbury Basin (Berndt et al., 2003; Gay et al., 2004; Gay and Berndt, 2007; Goulty, 2008). But that is not the only factor for the pockmark formation because 6 out of 11 pockmarks are not spatially consistent with the PF locations (Figure 5B). Other possibilities cannot be excluded such as bottom currents and differential compaction. In addition, PFs are often laterally impermeable, which may prevent lateral fluid migration and may form fault-lithologic traps when combined with nearby sand bodies (He et al., 2010; Andresen and Huuse, 2011).

Previous studies often regarded PFs as fluid migration pathways for shallower hydrocarbon or gas hydrate accumulations (Ding et al., 2013; Alrefaee et al., 2018; Velayatham et al., 2021). However, other fluid conduits cannot be ruled out such as the major faults, gas chimneys unconformities surface, or lateral migration along permeable layers to the structural highs. According to Ostanin et al. (2012), the widely distributed PFs found in Hammerfest Basin within Campanian layers may act as pathways for vertical hydrocarbon migration. However, the localized high amplitude anomalies are more likely to be distributed along the Palaeocene-Early Eocene Faults and the major deep 1^st order faults (Figure 6A). We propose that the hydrocarbon fluid from the deeper reservoirs may initially migrate along the 1^st order faults and then continue to migrate along the PEEFs to form the high amplitude anomalies. These enhanced amplitude anomalies may be due to the presence of hydrocarbons or diagenesis effect, both of which are related to the fluid migration (O'Brien et al., 2005). Velayatham et al. (2021) proposed that the PFs in the Ceduna Sub-basin, offshore South Australia may act as fluid migration pathways due to the spatial association of the PFs with shallower pipes (Figure 6B). However, after careful examination of the seismic profiles and plan view maps in the Ceduna Sub-basin, it seems insufficient to determine the fluids that escaped along the pipes were from underneath PFs which provided the fluids (Velayatham et al., 2021). Other possibilities of fluid supply should not be excluded such the migration along the major deep faults or lateral permeable sedimentary layers. Based on comprehensive analysis of enormous global PF case studies, we propose that in most PF cases the PFs play very limited roles in facilitating vertical long-distance fluid migration.

Many PFs occur in the fine-grained sediments layer of the Pearl River Mouth Basin in the northern South China Sea, but they may have played limited role in fluid migration (Sun et al., 2012; Yang et al., 2017). According to Sun et al. (2012) and Yang et al. (2017), the fluid migration system is mainly composed of gas chimneys and major large faults, and the enhanced seismic reflections representing shallow gas accumulation are mainly located on top of the gas chimneys (Figure 7). Meanwhile, the PF-bearing layer lacks local amplitude anomalies, and high amplitude gas reservoirs even exist underneath the PFs, both indicating that the PF-bearing layer should be impermeable at present and may hinder upward migration of the underlying fluids.

In the Lower Congo Basin, Gay et al. (2006) stated that the PFs allow deeper fluid migration to reach the seabed because of the upward deflection of the base of the gas hydrate stability zone (indicated by bottom-simulating reflector BSR in seismic), which was interpreted to be caused by a localized positive heat flow anomaly related to the PFs (Figure 8A). However, it has been noticed in this study that the deflection of the BSR solely appears around the chimney zone rather than throughout the whole PF-bearing layer. In the same area, Andresen and Huuse (2011) described the high reflectivity zone between the two PF-bearing layers which represents free gas accumulation. On the contrary, the vertical configuration of the free gas zone and the PF-bearing layer suggests the limited role of PFs as fluid migration conduits. It is more possible that the PF-bearing layer act as a seal trapping the free gas underneath. Additionally, localized high amplitude anomalies only exist along the gas chimney, indicating the positive role of gas chimney in fluid migration (Andresen and Huuse, 2011; Figure 8B,C). In this study, we interpret the high reflectivity zone between the two PF-bearing layers as sandstone-rich porous reservoir bodies which deposited as turbidite sandstone body in a background of fine-grained marine sediments. In this condition, good configuration of reservoir-seal led to the accumulation of free gas zone, with gas chimney being the major conduits transporting free gas from the deep. Considering the role of sandstone bodies as a barrier to PF propagation, the distribution and geometric attributes of PFs can be used to delineate sandstone bodies in the PF-bearing layer, which can be applied in date-poor areas (Jackson et al., 2014; Turrini et al., 2017).

The activity of the PFs is intermittent which can last over several millions of years, and therefore the role of PFs in fluids migration may be spatially and temporally variable (Berndt et al., 2003; Gay and Berndt, 2007; Berndt et al., 2012). The integrity of a fine-grained seal can be destroyed by the initial dewatering of host sediments, slip activity of the PFs, or the reactivation of PFs due to the contraction during tectonic rifting period, which results in the transient fluid leakage (Gay et al., 2006; Gay and Berndt, 2007; Hustoft et al., 2007; Seebeck et al., 2015). However, the PFs should be inactive and therefore probably impermeable in a larger geological time-scale. Additionally, the PF-bearing layers are often composed of primarily very fine-grained smectitic claystones or carbonate chalks, with generally low permeability (<10⁻¹⁷ m²) (Hibsch et al., 2003; Hansen et al., 2004; Tewksbury et al., 2014; Antonellini and Mollema, 2015).

It is widely acknowledged that the existence of PFs potentially increases the overall vertical permeability of fine-grained strata, slightly higher than the permeability of the host lithology (Evans et al., 1997; Gay et al., 2004; Faulkner et al., 2010; Ilg et al., 2012). However, this may not be large enough to allow fluid to flow smoothly through the PF-bearing layer. For example, the PFs in offshore Australia exist in the cap rock above the potential CO₂ reservoirs (Seebeck et al., 2015), suggesting the role of the PF-bearing layer as a seal. In contrast, the presence of PFs in coarse-grained sediments (e.g. porous sandstones) may reduce the permeability of PF-bearing layer, probably due to the deformation bands with cataclasis (Antonellini and Mollema, 2015). For example, the permeability of deformation band with cataclasis of the PFs is ∼10⁻¹⁵ m² at Arches National Park, while the permeability of deformation bands without cataclasis can reach 10⁻¹⁴ to 10⁻¹³ m² in the same area (Antonellini et al., 1994).

To summarize, the PFs developed in the sealing strata of fine-grained sediments may have improved the permeability for certain degree, but were still not favourable for fluid migration. Alternatively, the presence of PFs may even improve the sealing ability over time when clay smeared along the fault surfaces (Lonergan et al., 1998a; Turrini et al., 2017). In addition, Cartwright et al. (2007) proposed that the rate of fluid migration through PFs is relatively low on geological time scales, and therefore the presence of PFs alone cannot cause extended leakage from deeper hydrocarbon reservoirs. Generally, the role of PFs in fluid migration may be limited due to the PF-bearing layers act more as seals in most of the global PF cases.